Early experiences of integrating an artificial intelligence-based diagnostic decision support system into radiology settings: a qualitative study

This study examined the early experiences of stakeholders using an AI-based imaging software tool called Veye Lung Nodules (VLN) for detecting and classifying pulmonary nodules in chest CT scans. Interviews and observations were conducted with clinicians, decision-makers, suppliers, patients, and academics. The findings showed that clinicians found VLN easy to use with minimal disruption to workflow, but there were differences in usage patterns between experts and novices. Contextual variations and implementation challenges, such as integration with existing systems and data protection, were also identified. The study highlights the need for further research on the socio-organizational factors influencing the performance of diagnostic AI.

This study looked at how doctors and other stakeholders used a computer program called Veye Lung Nodules (VLN) to help them detect and classify lung nodules in chest scans. They found that most doctors found the program easy to use and it didn't disrupt their work too much. However, there were differences in how experienced doctors and less experienced ones used the program. They also found that the program worked differently in different hospitals and had some challenges with integrating it into existing systems and protecting patient data. Overall, they concluded that more research is needed to understand how these computer programs can be used effectively in healthcare.

There were differences in how experienced doctors and less experienced ones used the VLN program. They also found that the program worked differently in different hospitals and had some challenges with integrating it into existing systems and protecting patient data. Overall, they concluded that more research is needed to understand how these computer programs can be used effectively in healthcare.